Spaces:
Sleeping
Sleeping
File size: 14,583 Bytes
e0aa230 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 |
"""
Query Processor Module
This module is responsible for processing user queries and converting
them to vector embeddings for retrieval.
Technologies: Gemini Embedding v3, LangChain, Pinecone
"""
import logging
import time
from typing import Dict, List, Any, Optional
from datetime import datetime, timedelta
class QueryProcessor:
"""
Processes user queries and converts them to vector embeddings.
Features:
- Query preprocessing and normalization
- Query embedding generation
- Context retrieval from vector database
- Query expansion and caching
- Metadata filtering and ranking
"""
def __init__(
self, embedding_generator, vector_db, config: Optional[Dict[str, Any]] = None
):
"""
Initialize the QueryProcessor with dependencies.
Args:
embedding_generator: Instance of EmbeddingGenerator
vector_db: Instance of VectorDB
config: Configuration dictionary with processing parameters
"""
self.embedding_generator = embedding_generator
self.vector_db = vector_db
self.config = config or {}
self.logger = logging.getLogger(__name__)
# Configuration settings
self.top_k = self.config.get("top_k", 5)
self.similarity_threshold = self.config.get("similarity_threshold", 0.7)
self.max_context_length = self.config.get("max_context_length", 4000)
self.enable_caching = self.config.get("enable_caching", True)
self.cache_ttl = self.config.get("cache_ttl", 3600) # 1 hour
# Query cache and history
self.query_cache = {}
self.query_history = []
self.logger.info("QueryProcessor initialized with advanced features")
def process_query(
self, query: str, filter: Optional[Dict[str, Any]] = None
) -> Dict[str, Any]:
"""
Process a user query and retrieve relevant context.
Args:
query: User query string
filter: Optional metadata filter for search
Returns:
Dictionary containing query, retrieved context, and metadata
"""
if not query or not query.strip():
return {
"query": query,
"context": [],
"total_results": 0,
"error": "Empty query provided",
}
self.logger.info(f"Processing query: {query[:100]}...")
start_time = time.time()
try:
# Check cache first
cache_key = self._generate_cache_key(query, filter)
if self.enable_caching and cache_key in self.query_cache:
cached_result = self.query_cache[cache_key]
if self._is_cache_valid(cached_result["timestamp"]):
self.logger.info(" Returning cached result")
cached_result["from_cache"] = True
return cached_result
# Preprocess the query
processed_query = self._preprocess_query(query)
expanded_queries = self._expand_query(processed_query)
# Generate embeddings for all query variations
all_results = []
for q in expanded_queries:
query_embedding = self.embedding_generator.generate_query_embedding(q)
if query_embedding:
# Search for similar vectors
search_results = self.vector_db.search(
query_embedding=query_embedding,
top_k=self.top_k * 2, # Get more results for better filtering
filter=filter,
)
all_results.extend(search_results)
# Deduplicate and rank results
unique_results = self._deduplicate_results(all_results)
ranked_results = self._rank_results(unique_results, query)
# Filter results by similarity threshold
filtered_results = [
result
for result in ranked_results[: self.top_k]
if result.get("score", 0) >= self.similarity_threshold
]
# Extract and format context
context = self._extract_context(filtered_results)
# Prepare result
result = {
"query": query,
"processed_query": processed_query,
"expanded_queries": expanded_queries,
"context": context,
"total_results": len(filtered_results),
"processing_time": time.time() - start_time,
"timestamp": datetime.now(),
"from_cache": False,
}
# Cache the result
if self.enable_caching:
self.query_cache[cache_key] = result.copy()
# Add to query history
self._add_to_history(query, len(filtered_results))
self.logger.info(f"Query processed in {result['processing_time']:.2f}s")
return result
except Exception as e:
self.logger.error(f"❌ Error processing query: {str(e)}")
return {
"query": query,
"context": [],
"total_results": 0,
"error": str(e),
"processing_time": time.time() - start_time,
}
def _preprocess_query(self, query: str) -> str:
"""
Preprocess the query for better embedding generation.
Args:
query: Raw query string
Returns:
Preprocessed query string
"""
# Remove extra whitespace
query = " ".join(query.split())
# Remove special characters that might interfere
import re
query = re.sub(r"[^\w\s\-\?\!]", " ", query)
# Normalize question words
question_words = {
"whats": "what is",
"hows": "how is",
"wheres": "where is",
"whos": "who is",
"whens": "when is",
}
for abbrev, full in question_words.items():
query = query.replace(abbrev, full)
return query.strip()
def _expand_query(self, query: str) -> List[str]:
"""
Expand the query with variations for better retrieval.
Args:
query: Preprocessed query
Returns:
List of query variations
"""
expanded = [query]
# Add question variations
if not any(
q in query.lower() for q in ["what", "how", "why", "when", "where", "who"]
):
expanded.append(f"what is {query}")
expanded.append(f"how does {query} work")
# Add definition variation
if "definition" not in query.lower() and "define" not in query.lower():
expanded.append(f"{query} definition")
# Add example variation
if "example" not in query.lower():
expanded.append(f"{query} examples")
return expanded[:3] # Limit to 3 variations
def _deduplicate_results(
self, results: List[Dict[str, Any]]
) -> List[Dict[str, Any]]:
"""
Remove duplicate results based on content similarity.
Args:
results: List of search results
Returns:
Deduplicated results
"""
seen_ids = set()
unique_results = []
for result in results:
result_id = result.get("id")
if result_id and result_id not in seen_ids:
seen_ids.add(result_id)
unique_results.append(result)
return unique_results
def _rank_results(
self, results: List[Dict[str, Any]], query: str
) -> List[Dict[str, Any]]:
"""
Rank results based on multiple factors.
Args:
results: List of search results
query: Original query
Returns:
Ranked results
"""
query_words = set(query.lower().split())
for result in results:
# Base score from similarity
base_score = result.get("score", 0.0)
# Boost score based on text relevance
text = result.get("metadata", {}).get("text", "").lower()
text_words = set(text.split())
word_overlap = len(query_words.intersection(text_words))
relevance_boost = word_overlap / max(len(query_words), 1) * 0.1
# Boost score based on source type
source = result.get("metadata", {}).get("source", "")
source_boost = 0.0
if source.endswith(".pdf"):
source_boost = 0.05 # PDFs often contain structured info
elif "http" in source:
source_boost = 0.02 # Web content
# Calculate final score
final_score = base_score + relevance_boost + source_boost
result["final_score"] = min(final_score, 1.0)
# Sort by final score
return sorted(results, key=lambda x: x.get("final_score", 0), reverse=True)
def _extract_context(
self, search_results: List[Dict[str, Any]]
) -> List[Dict[str, Any]]:
"""
Extract and format context from search results.
Args:
search_results: List of search results from vector database
Returns:
List of formatted context items
"""
context = []
total_length = 0
for result in search_results:
# Extract text content from metadata
text = result.get("metadata", {}).get("text", "")
# Check if adding this context would exceed the limit
if total_length + len(text) > self.max_context_length and context:
break
# Format context item with enhanced metadata
context_item = {
"text": text,
"score": result.get("score", 0),
"final_score": result.get("final_score", result.get("score", 0)),
"source": result.get("metadata", {}).get("source", "unknown"),
"chunk_id": result.get("id", ""),
"metadata": result.get("metadata", {}),
"relevance_rank": len(context) + 1,
}
context.append(context_item)
total_length += len(text)
self.logger.info(
f"Extracted {len(context)} context items (total length: {total_length})"
)
return context
def _generate_cache_key(self, query: str, filter: Optional[Dict[str, Any]]) -> str:
"""Generate a cache key for the query."""
import hashlib
filter_str = str(sorted(filter.items())) if filter else ""
cache_string = f"{query.lower().strip()}{filter_str}"
return hashlib.md5(cache_string.encode()).hexdigest()
def _is_cache_valid(self, timestamp: datetime) -> bool:
"""Check if cached result is still valid."""
return datetime.now() - timestamp < timedelta(seconds=self.cache_ttl)
def _add_to_history(self, query: str, result_count: int):
"""Add query to history for analytics."""
self.query_history.append(
{
"query": query,
"timestamp": datetime.now(),
"result_count": result_count,
}
)
# Keep only last 100 queries
if len(self.query_history) > 100:
self.query_history = self.query_history[-100:]
def get_query_suggestions(self, partial_query: str) -> List[str]:
"""
Generate query suggestions based on partial input and history.
Args:
partial_query: Partial query string
Returns:
List of suggested queries
"""
suggestions = []
# Add suggestions from query history
for hist_item in reversed(self.query_history[-20:]): # Last 20 queries
hist_query = hist_item["query"]
if (
partial_query.lower() in hist_query.lower()
and hist_query not in suggestions
):
suggestions.append(hist_query)
# Add template-based suggestions
if len(suggestions) < 3:
templates = [
f"What is {partial_query}?",
f"How does {partial_query} work?",
f"Examples of {partial_query}",
f"{partial_query} definition",
f"{partial_query} best practices",
]
for template in templates:
if template not in suggestions:
suggestions.append(template)
if len(suggestions) >= 5:
break
return suggestions[:5]
def get_query_analytics(self) -> Dict[str, Any]:
"""
Get analytics about query patterns.
Returns:
Dictionary with query analytics
"""
if not self.query_history:
return {"total_queries": 0, "cache_hit_rate": 0.0}
total_queries = len(self.query_history)
recent_queries = [q["query"] for q in self.query_history[-10:]]
# Calculate average results per query
avg_results = sum(q["result_count"] for q in self.query_history) / total_queries
# Most common query patterns
query_words = []
for q in self.query_history:
query_words.extend(q["query"].lower().split())
from collections import Counter
common_words = Counter(query_words).most_common(5)
return {
"total_queries": total_queries,
"average_results_per_query": round(avg_results, 2),
"recent_queries": recent_queries,
"common_query_words": common_words,
"cache_size": len(self.query_cache),
}
def clear_cache(self):
"""Clear the query cache."""
self.query_cache.clear()
self.logger.info("Query cache cleared")
def clear_history(self):
"""Clear the query history."""
self.query_history.clear()
self.logger.info("Query history cleared")
|